ABSTRACT This is a competitive renewal of a longitudinal study of late-life suicide. Our ultimate aim is the greater understanding of decision processes that underlie suicidal behavior in old age. Toward this goal, we have accrued a high-risk cohort of depressed older adults (including 173 with past suicide attempts and 82 with serious suicidal ideation) with prospectively ascertained high rates of suicidal behavior. AIM 1. To date, we have shown robust associations between decision process alterations and cognitive control deficits on one hand, and the history of suicidal behavior, on the other, consistent with the notion that these impairments undermine the search for constructive alternatives to suicide. As a crucial next step, we will confirm these associations prospectively through prediction of incident suicide attempts and death by suicide. We predict that decision indices will have predictive power after accounting for recent suicidal ideation. To identify the most useful predictors of incident suicidal behavior, we will also develop a parsimonious prediction model based on a broad range of potential risk factors. AIM 2. Our studies to date have focused on decision-making in non-social contexts. However, many precipitants and deterrents to suicide are rooted in social relationships. In a suicidal crisis, people often fail to consider how devastating their suicide would be for those around them. Normative age-related events affecting social status, such as disability or retirement, may trigger suicidal behavior in late life in a predisposition- dependent manner. To probe the mechanisms underlying catastrophic decisions in response to loss of social dominance, we propose a study of social decision-making involving a loss-of-status manipulation. In addition, we propose to test whether blunted motivational effects of social relationships in high-suicide-risk individuals are related to alterations in Pavlovian behavioral biases and neural responses evoked by social stimuli. To achieve these aims, we propose to extend the follow-up of the current cohort and expand it with 80 newly recruited attempters and 60 ideators. Our study of social decision-making will employ a case-control design (including depressed individuals with no history of suicide attempt or ideation and non-psychiatric controls) to detect effects beyond those by conferred by depression and suicide ideation, and to minimize confounders. We will use a decision neuroscience framework, combining model-based imaging with multilevel clinical, behavioral, and cognitive assessments. We have established procedures to conduct this research in a safe, ethical manner. The investigative team contributes complementary expertise in suicidal behavior and neurocognitive processes (Szanto), in Pavlovian and instrumental reward learning and model-based imaging (Dombrovski, Hallquist), social decision-making (Crockett), analysis of neural and longitudinal data (Galfalvy), decision neuroscience and experimental design (Clark), neuropsychology of geriatric depression (Butters).